How I Built RevisionLab: Lessons From a Revision Startup
RevisionLab started with a simple, frustrating observation: students were working hard but revising in ways that did not work. Bright, motivated young people were re-reading notes for hours and still walking into exams underprepared. This is the story of why I built RevisionLab, the lessons learned along the way, and what they might offer anyone thinking of building something in education.
The problem I kept seeing
Again and again, the issue was not effort or ability; it was method. Students equated revision with re-reading and highlighting, the very techniques research shows are weakest. Meanwhile the methods that genuinely work, active recall and spaced repetition, were rarely taught and even more rarely built into how students actually studied. There was a gap between what the science says and what students do.
The idea: make good technique the default
RevisionLab's core idea was to stop relying on students to know and apply the right techniques, and instead build those techniques into a tool. If a platform organised your subjects, scheduled topics on a spaced timetable, and turned your notes into active recall practice automatically, then doing the effective thing would become the easy, default thing. Good revision should not depend on willpower or knowing the theory.
Lesson 1: Solve a real, specific problem
The most useful starting point was a problem I understood deeply and personally. Building for a vague audience leads to a vague product. Focusing on a specific, painful, real problem, students revising ineffectively, gave every decision a clear test: does this help someone revise better? If not, it did not belong.
Lesson 2: Start smaller than feels comfortable
It is tempting to try to build everything at once. The better path was to start with the single most valuable thing, the core revision experience, and resist the urge to bolt on features. A focused product that does one thing well beats a sprawling one that does many things poorly, and it is far easier to improve.
Lesson 3: Ground it in evidence, not opinion
Education is full of myths, like learning styles, that feel intuitive but do not hold up. Anchoring RevisionLab in well-supported cognitive science, rather than what merely felt right, kept the product genuinely useful. When building in any serious field, it pays to respect the evidence over assumptions.
Lesson 4: Content and product reinforce each other
Helping students with free, genuinely useful content, like the articles on this blog, does two things: it helps people directly, and it builds trust and an audience for the product. The blog and the tool are not separate; teaching good revision and providing a tool that applies it are two sides of the same mission.
Lesson 5: Use modern tools, keep human judgement
Modern AI tools make it possible to build and create faster than ever, from drafting content to prototyping features. But in education, accuracy and trust are everything. The lesson was to use these tools to move quickly while keeping human expertise firmly in charge of quality. Speed is worthless if the output cannot be trusted.
Where it goes from here
RevisionLab is still evolving, and the mission stays the same: close the gap between how students revise and how they should. Every feature is judged against one question, does it help a student learn more effectively? That north star keeps the project honest as it grows.
Frequently asked questions
What is RevisionLab?
RevisionLab turns the most effective revision techniques, active recall and spaced repetition, into a tool that organises your subjects and turns your notes into structured, spaced practice.
Why was RevisionLab created?
Because students were working hard but using weak revision methods. The aim was to build effective techniques into a tool so good revision becomes the default.
What are the main lessons from building it?
Solve a real, specific problem; start small and focused; ground the product in evidence; let content and product reinforce each other; and use modern tools without giving up human judgement.
Do I need to be technical to build something in education?
Less than before. No-code and AI tools lower the barrier, but deep understanding of the problem and a commitment to accuracy matter most.
RevisionLab exists to make effective revision the easy default. If this blog has helped you revise better, that is the same mission in action, and the tool simply takes it further.
What I actually built
RevisionLab developed into a set of private study hubs rather than a public library of generic worksheets. A hub organises a learner's subjects, current topics, useful resources and next tasks in one place. The practical goal is to reduce the question students often face at the start of a session: “What should I do now?”
The most useful parts are deliberately simple. A student needs a clear next action, a way to practise without immediately seeing the answer, and a record of what still needs work. A large content library can look impressive while making those decisions harder.
Three product lessons from the work
1. The next action matters more than the dashboard
Early ideas tended to add more categories, pages and resources. In practice, every extra choice creates another place for a learner to hesitate. I now prefer a short, specific instruction such as “complete these five algebra questions, mark them, and retry the two errors” over a dashboard full of vague progress indicators.
2. Feedback loops beat passive content
A revision resource becomes useful when it produces evidence. That evidence might be a score, an error pattern, a question the student could not explain, or a topic that needs returning to next week. This is why the core RevisionLab workflow is based on attempting, checking, correcting and retrying rather than simply collecting notes.
3. Privacy is part of the product
Student hubs are private. Public articles can discuss revision methods, but names, personal progress and individual learning records do not belong in public examples. Keeping that boundary clear affects how pages are designed, how access is handled and which screenshots or stories can be shared.
What I would do differently
I would publish fewer features and test each one against a real study session sooner. It is easy to spend time polishing a page that does not change what a learner does next. A smaller tool that helps someone identify and correct one misunderstanding is more valuable than a larger one that only looks complete.
I also would not treat “engagement” as the final measure. More time on a revision site is not automatically better. A successful session might be short: the learner identifies a gap, practises it, checks the result and leaves with a clear follow-up task.
Where the project is now
RevisionLab remains a working education project, and the blog is where I document the revision principles behind it. You can visit RevisionLab or use the Contact page to report a problem or suggest a topic.
Updated 13 August 2026 to add implementation detail and editorial context.
Comments
Post a Comment